In response to “significant confusion” in the bioinformatics community over conditions of software availability, the International Society for Computational Biology recently issued a number of recommendations via a statement on its website (www.iscb.org).
At the core of its proposal is a five-level structure identifying a range of licensing, pricing, and redistribution possibilities for bioinformatics software. “Level 0” software is available free of charge for educational institutions and non-profit research groups and for a possible fee to commercial institutions without the right to further redistribute the software; “Level 1” is free of charge to all institutions without the right of redistribution; “Level 2” is available free of charge to educational groups and non-profits along with source code, but without the right to redistribution; “Level 3” is free of charge to all institutions with source code; and “Level 4” provides the software free of charge to all institutions and individuals with source code and unlimited redistribution rights.
The definitions were developed with the hope of providing a standard reference for what many in the field see as increasingly muddled territory. Peter Karp, director of SRI’s Bioinformatics Research Group and a member of the ISCB panel that drafted the recommendations, noted that even seasoned software developers on the committee “didn’t really understand all the nuances of open source.”
The term has “diverged” to mean different things to different people, Karp said, and “different groups have tweaked it to mean what they want it to mean.”
The ISCB statement weighed in on a controversial topic in the open source community, noting that “government agencies that fund bioinformatics research should not require that software produced with government research funds must be distributed under open-source license (particularly given the ambiguity in the meaning of that term).”
“Our message is that because of the complexity in developing software, you can’t legislate one model for all cases,” said Karp. “You need to consider the particulars of a given research project.” However, he said, the ISCB agreed it was necessary to lay down a minimum level of availability for federally funded software development, which spurred the development of a formal set of recommendations.
The ISCB’s position on federally funded software is in line with that of the Open Bioinformatics Foundation, an umbrella group that supports the bio* projects. Unlike some more extreme open source advocacy groups, the OBF holds that it is up to software developers rather than funding agencies to determine the terms of availability for their work. Although OBF board member Ewan Birney of the European Bioinformatics Institute questioned the ISCB’s assessment of the level of misunderstanding about open source in the bioinformatics community, he said the society’s five-level system seemed to be a fair way of describing the options available to developers.
While the guidelines were intended to aid researchers and funding agencies in the grant-writing process, Karp said the ISCB is currently in discussions with the editors of Bioinformatics regarding use of the definitions in papers submitted to the journal as well.
In addition, the ISCB intends to host a discussion page on the topic on its website.